Hugging Face challenges DeepSeek: Project Open-R1 reveals secrets of Chinese AI
Hugging Face challenges DeepSeek: Project Open-R1 reveals secrets of Chinese AI The Hugging Face team presented the first results of the Open-R1 project aimed at reproducing the technologies of Chinese artificial intelligence DeepSeek-R1. Within a week, researchers managed to achieve significant progress in understanding and replicating this advanced system.
A key achievement was the successful reproduction of test results on the MATH-500 benchmark. Researchers confirmed impressive performance of various model versions: DeepSeek-R1-Distill-Qwen-32B achieved 95.0% accuracy compared to the claimed 94.3%, while the Llama-70B-based version showed 93.4% versus the official 94.5%.
During the study, a unique feature of DeepSeek-R1 was discovered – unprecedented length of generated responses. Analysis of distribution in the OpenThoughts dataset showed that the average response length is about 6000 tokens, and in some cases exceeds 20,000 tokens. “Considering that an average page contains approximately 500 words, and one token is slightly shorter than a word, many responses exceed 10 pages in volume,” researchers note.
To ensure research transparency, the Hugging Face team created an open Open-R1 leaderboard where the community can track progress in reproducing results. Special attention is paid to the issue of significant GPU memory requirements during training due to the need to generate long sequences.
The Open-R1 project, launched just a week ago, combined the efforts of various teams and the developer community. The main goal remains to reproduce the training pipeline and synthetic data of DeepSeek-R1, which will help better understand the operating principles of this advanced artificial intelligence system.
This initiative demonstrates a growing trend towards openness and collaboration in AI, where even the most complex technological achievements become the subject of collective study and reproduction by the global developer community.
Autor: AIvengo
For 5 years I have been working with machine learning and artificial intelligence. And this field never ceases to amaze, inspire and interest me.
Qualcomm welcomes TSMC's $100 billion investmentTaiwan Semiconductor Manufacturing Co. (TSMC)'s $100 billion investment in expanding production in the United States is "great news," said Qualcomm CEO Cristiano Amon in an interview with CNBC on Tuesday, March 4, 2025. According to him, this contributes to the diversification of semiconductor manufacturing locations.
DuckDuckGo strengthens generative AI integrationThe privacy-focused search service DuckDuckGo continues to strengthen its position in the field of generative artificial intelligence. According to a blog post published on Thursday, March 6, 2025, the company announced the completion of beta testing for its chat interface, which is now officially called Duck.ai, abandoning the more cumbersome name DuckDuckGo AI Chat.
Digital scandal at Los Angeles TimesBillionaire and Los Angeles Times owner Patrick Soon-Shiong, who introduced a new AI tool for generating opposing perspectives to opinion section materials, was unaware that the system created pro-KKK arguments less than 24 hours after launch — and even hours after the scandalous AI comments were removed from the publication's website. The incident created a huge obstacle for the Times, which seeks to bring back old subscribers and attract new ones through innovative technological solutions.
Google Shopping launches AI toolGoogle announced the launch of a new AI tool for the Shopping tab that will help users find clothing based on their verbal description. The announcement, made on Wednesday, March 5, 2025, also includes expanding the capabilities of augmented reality (AR) tools for cosmetics and virtual try-on.
"Cannot help with answers about elections and political figures": GeminiTechnology giant Google continues to limit the capabilities of its AI assistant Gemini in the area of political discourse, despite the fact that the company's main competitors, including OpenAI, Anthropic, and Meta, have already adapted their chatbots to discuss politically sensitive topics in recent months.
